Currently I was reading ISO3951-1 2022 version : Specification for single sampling plans indexed by acceptance quality limit (AQL) for lot-by-lot inspection for a single quality characteristic and a single AQL. I was wondering how ISO decided code letter for each batch size. And how this batch size determine the suggested different sample size and K value.
For me, K value is based on the tolerance interval that decided the coverage of population. It depends on the Producer's risk (alpha), Customer's risk (beta), AQL, RQL. And in a series of solving the combination of (Alpha, AQL), (Beta, RQL) to get the sample size and K value.
So here is my question :
1. How ISO decide how many sample we need in different batch size? Is there any statistical reason behind that?
2. How ISO calculate K? In my opinion, if we need to know at least we need to know confidence (for example : alpha) and coverage of population (for example : 1- AQL) to calculate constant K.
Maybe I miss some content in ISO3951-1 2022 or another standard. However, I am quite confused with batch size and acceptance sampling.
For me, K value is based on the tolerance interval that decided the coverage of population. It depends on the Producer's risk (alpha), Customer's risk (beta), AQL, RQL. And in a series of solving the combination of (Alpha, AQL), (Beta, RQL) to get the sample size and K value.
So here is my question :
1. How ISO decide how many sample we need in different batch size? Is there any statistical reason behind that?
2. How ISO calculate K? In my opinion, if we need to know at least we need to know confidence (for example : alpha) and coverage of population (for example : 1- AQL) to calculate constant K.
Maybe I miss some content in ISO3951-1 2022 or another standard. However, I am quite confused with batch size and acceptance sampling.
